ViRA: Toward Ultra-High-Resolution Visual Agents via Cross-Input Relative Policy Optimization A paper submitted to arXiv on 16 Nov 2026 proposes Cross-Input Relative Policy Optimization (CrossRPO), a reinforcement learning method that trains visual agents on contrast sets — one question asked about several versions of an image that differ only in a small evidence region with different answers — so that only reading the evidence solves the whole set. The authors report CrossRPO reaches 63.3% set accuracy on ViRA-Bench at maximum resolution versus 59.3% for GRPO, using contrast sets built from 21- to 120-megapixel images and training ViRA-9B and ViRA-27B on only about 3,500 ViRA-Train items each. The authors state ViRA-9B is the most accurate open model of its size on all benchmarks and ViRA-27B outperforms Claude Haiku 5.5 and GPT-6-Luna, both at their highest thinking level, on most benchmarks. Computer Science Computer Vision and Pattern Recognition Submitted on 16 Nov 2026 Title:ViRA: Toward Ultra-High-Resolution Visual Agents via Cross-Input Relative Policy Optimization View PDF http://arxiv.org/pdf/2611.21605v1 HTML experimental https://arxiv.org/html/2611.21605v1 Abstract:Visual agents answer questions about ultra-high-resolution photos by zooming into the relevant regions. However, existing agents often answer from common sense instead of reading the evidence. Reinforcement learning RL on the final answer cannot correct this, because a guess earns the same reward as reading. We propose Cross-Input Relative Policy Optimization CrossRPO , which trains agents on contrast sets. A contrast set asks one question about several versions of an image that differ only in a small region with the evidence and have different answers, so only reading the evidence solves the whole set. CrossRPO rewards solving the whole set, and we prove that it does not reinforce a set answered from common sense. We build contrast sets for training ViRA-Train and evaluation ViRA-Bench by editing one small object in images of 21 to 120 megapixels, and we photograph further sets with real changes ViRA-Bench-Real to exclude editing traces. We train two Visual Reasoning Agents, ViRA-9B and ViRA-27B, with CrossRPO on only about 3,500 items of ViRA-Train each. CrossRPO achieves 63.3% set accuracy on ViRA-Bench at the maximum resolution, against 59.3% for GRPO. ViRA-9B is the most accurate open model of its size on all benchmarks, and ViRA-27B outperforms Claude Haiku 5.5 and GPT-6-Luna, both at their highest thinking level, on most benchmarks. Additional Features References & Citations Loading... Bibliographic and Citation Tools Bibliographic Explorer What is the Explorer? https://info.arxiv.org/labs/showcase.html arxiv-bibliographic-explorer Connected Papers What is Connected Papers? https://www.connectedpapers.com/about Litmaps What is Litmaps? https://www.litmaps.co/ scite Smart Citations What are Smart Citations? https://www.scite.ai/ Code, Data and Media Associated with this Article alphaXiv What is alphaXiv? https://alphaxiv.org/ CatalyzeX Code Finder for Papers What is CatalyzeX? https://www.catalyzex.com DagsHub What is DagsHub? https://dagshub.com/ Gotit.pub What is GotitPub? http://gotit.pub/faq Hugging Face What is Huggingface? https://huggingface.co/huggingface ScienceCast What is ScienceCast? https://sciencecast.org/welcome Demos Recommenders and Search Tools Influence Flower What are Influence Flowers? https://influencemap.cmlab.dev/ CORE Recommender What is CORE? https://core.ac.uk/services/recommender arXivLabs: experimental projects with community collaborators arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website. Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them. Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs https://info.arxiv.org/labs/index.html .